Anthropic has updated its privacy policy — and tucked inside is a change that will touch a lot of users. Starting July 8, 2026, the company reserves the right to run age and identity verification on Claude users. In plain terms: upload a photo ID, plus a live selfie for matching.
Who’s affected?
The rule targets the consumer tiers — Claude Free, Pro, and Max. Team, Enterprise, and Platform customers are explicitly excluded. The checks don’t run through Anthropic itself but through outside providers: age verification is handled by Yoti, while identity checks go through Persona. According to Anthropic, ID and selfie images are stored by Persona, not on Anthropic’s own systems.
Anthropic’s rationale
The framing is worth noting. The company calls this a voluntary step that no current law requires. The reasoning is about responsibility — as Claude becomes more powerful and more agentic, knowing who’s operating the tool matters more. At the same time, Anthropic reaffirms its familiar privacy commitments: user data isn’t sold, Claude stays ad-free, and the opt-out for model training remains in place.
My take
This is one of those announcements where two things grind against each other. On one side, the safety logic checks out: as AI agents increasingly act on their own, identity and accountability stop being a footnote. On the other, “upload your ID and a selfie” is a steep ask — especially for a product you could simply open in a browser until now.
Routing the data to third parties instead of Anthropic is a tidy privacy move, but it doesn’t change the fact that more biometric data gets collected than before. The interesting part is how this plays out in practice: “reserves the right” isn’t the same as “requires by default.” Most likely the check only kicks in on certain triggers. Either way, if you use Claude personally, put July 8 on your calendar.
Sources: The Register, Cybernews, Medianama